Short answer
Talent intelligence tools serve different decisions: recruiting candidates, opening internal opportunities, interpreting external labour markets, planning workforce supply, or adding current employee accounts to a human review. Choose the category from the decision you need to make, then write a brief that names the inputs, their provenance, correction route, uncertainty, human owner, implementation dependency and exit criterion.
If you already know the required category and need supplier names, go to the talent intelligence platform comparison and vendor scorecard. Use the brief below to decide which products belong on that shortlist.
A product can be sophisticated and still be the wrong category for the job. A talent marketplace will not by itself settle a capacity scenario. A labour-market data set will not explain a local handover problem. A conversation tool should not be bought as a skills inventory or candidate-matching engine.
Choose the category from the decision
Start by finishing this sentence: “By [date], [named owner] needs to decide [specific action] for [defined population or work], using [named sources].” If the sentence cannot be written, a vendor comparison will reward broad claims rather than workflow fit.
| Decision to make | Useful inputs | Category to evaluate | Output to inspect | Category boundary |
|---|---|---|---|---|
| Fill a defined vacancy | Role requirements, candidate profiles, recruiter judgement | Recruitment matching | A traceable candidate shortlist | It does not decide who to hire |
| Offer roles, projects or development routes internally | Opportunity rules, employee interests, role and skills context | Internal opportunities or talent marketplace | Eligible opportunities and an explainable route to them | It does not prove capability from a profile alone |
| Understand external availability or changing occupations | Public labour-market and occupational data | Labour-market intelligence | A documented market view or assumption | It does not describe work inside your organisation |
| Compare supply, demand, cost and capability scenarios | Business assumptions, workforce records, financial constraints | Workforce planning | A challengeable scenario | It does not replace the systems that execute the plan |
| Prepare a conversation about current work | Employee accounts, current work examples, manager context | Guided employee conversations | A focused brief and human-owned next action | It is not a talent marketplace, skills inventory or automatic match |
Categories can appear in one suite, and a programme may use more than one. The table is a commissioning aid, not a market taxonomy. For example, Eightfold's product overview presents distinct acquisition, talent-management, workforce-exchange and resource-management products. Workday's talent-management overview describes a suite that includes talent, skills and internal-mobility work. Those official descriptions identify the products vendors offer; they do not prove that a particular configuration fits your decision.
Write the buying brief before the demo
Give every prospective supplier the same brief. It turns “show us your AI” into a testable request.
| Brief field | Prompt for the buying team |
|---|---|
| Decision | What decision will change after the output is reviewed? Name one decision, not a transformation programme. |
| Input and provenance | Which records, employee accounts, taxonomies or external data enter? For each, who owns it, when was it updated and what does it actually represent? |
| Correction route | Who can challenge an outdated profile, disputed project, incorrect role requirement or missing account, and how will the correction appear downstream? |
| Uncertainty | What should remain marked as missing, declared, inferred, disputed or unverified rather than presented as fact? |
| Human owner | Who reads the output, decides the next action and records the decision in the existing process? |
| Dependencies | Which HRIS, opportunity rules, taxonomy, permissions, local employment process, data owner or change capacity must exist before the workflow can work? |
| Exit criterion | What result would make the team stop, narrow or defer the purchase? Examples: source cannot be traced, corrections cannot be demonstrated, the owner cannot use the output, or a required system dependency is unavailable. |
This brief asks for transparency without assuming that every tool needs the same data. The US Department of Labor-sponsored O*NET database provides occupation, task and skills information that can help a US organisation establish shared language. It is an occupational reference, not a record of what an individual employee can do in a particular UK or US workplace.
For AI-assisted outputs, include a failure test in the brief. The NIST AI Risk Management Framework is a useful public structure for documenting context, governance, measurement and ongoing management. It does not certify a vendor. Ask the vendor to show what happens when a source is old, a profile is missing, an account is contested or a user asks for an output outside the agreed purpose.
A fictional US and UK commissioning case
Northline Services is fictional. It runs customer-support teams in the US and the UK and plans to introduce a new support product in six months. Its People director initially asks for “a talent intelligence platform.” The group rewrites that request into two separate decisions.
First, the workforce-planning lead must decide whether to hire, develop or redeploy support specialists. This requires demand assumptions, role definitions, current workforce records and Finance input. The team commissions a planning-category evaluation and names the planning lead as owner.
Second, two UK teams report that a new escalation handover is difficult to use. This is not a question about internal matching. The local manager needs current examples of the work and a human conversation about one practical change. The team commissions a guided-conversation workflow with a people partner as reviewer.
The brief says that a role title, training completion and an employee's own account are different source types. A stale training record must remain stale; an account must not become a verified capability merely because it appears in a tool. Employees need a route to correct relevant context. The manager needs a brief for the discussion, not raw employee conversations. The outcome is a documented next action and date for review, not a prediction of performance, retention or mobility.
The team stops the conversation workflow if it cannot make the manager brief's source boundaries clear, or if no suitable review owner can commit to the next step. It stops the planning evaluation if the supplier cannot expose the assumptions and dependencies behind a scenario. Those are useful exit criteria because they test the proposed workflow, not a sales presentation.
Where Lontra fits in the brief
Lontra fits the narrower current-work category. Invited employees describe a work situation through guided conversations. Managers receive briefs without raw employee responses; people then hold the conversation and decide the next action. Where permitted, dated context can be revisited in a later campaign, but it remains context to check rather than a verified fact or an automatic people decision.
This can complement a talent, mobility or planning workflow when a named owner needs current work examples before a human discussion. It does not provide a complete skills inventory, a talent marketplace, labour-market intelligence, automated candidate matching or a staffing decision. See the current product workflow if employee conversations are the job you are commissioning.
The Lontra trial is one campaign with up to 30 invitations over 60 days, with no card required. Use it to test one question, a named review owner, the brief boundary and a dated follow-up. Studio is a separate paid subscription module, activated with guidance for creating human-reviewed learning material from an approved practice. It is not required for this conversation workflow.
Use the right next guide
This article is for choosing a category and writing the brief. Once the category is chosen, use the talent intelligence platform guide to plan the commissioning and pilot records. Use the talent intelligence platform comparison when you need an equal-scenario vendor scorecard. Bring the same buying brief to each stage so suppliers are evaluated against the decision you actually need to make.
Sources
- O*NET Database
- NIST AI Risk Management Framework
- Eightfold product overview
- Workday Talent Management overview
Frequently asked questions
What are talent intelligence tools?
Talent intelligence tools are products used for different workforce decisions, including recruitment matching, internal opportunities, labour-market intelligence, workforce planning, and current employee context. The right category depends on the decision, sources and human workflow required.
How do I choose a talent intelligence tool category?
Start with one decision, then specify the required inputs, provenance, correction route, uncertainty, human owner, implementation dependency and exit criterion. Do not choose from a feature list alone.
Can a conversation tool replace a talent intelligence platform?
Usually no. Guided conversations can add current employee accounts to a human review, while talent-intelligence platforms may serve recruitment, mobility, skills or planning workflows. Evaluate each against its stated job.